CohereTranslation site

Company Overview

Cohere Inc. is an enterprise-level big modeling (LLM) company founded in 2019 and headquartered in Toronto, Canada, with offices in San Francisco, London, and New York. Co-founded by former Google Brain researcher Aidan Gomez (current CEO and co-inventor of the Transformer architecture), Nick Frosst and Ivan Zhang.

creator: Aidan Gomez, Ivan Zhang, and Nick Frosst, with Aidan Gomez being one of the lead authors of the Attention Is All You Need paper and having worked on the Google Brain team.

Financing::

• In September 2021, closed a $40 million Series A round led by Index Ventures, Section 32 and Radical Ventures.
• In February 2022, it secured $125 million in Series B financing led by Tiger Global.
• In June 2023, closed a $270 million Series C round led by Inovia Capital with participation from NVIDIA, Oracle and Salesforce.
• In July 2024, a $500 million Series D financing round was completed, led by Canadian pension fund PSP Investments, valuing the company at $5.5 billion.

Core Products and Services

• Language Modeling API Service: Enterprises can quickly access powerful natural language processing capabilities through Cohere's APIs for document management, content generation and semantic search scenarios.

• SaaS SolutionsFor example, the Coral Knowledge Assistant can help employees within an organization to quickly retrieve information, automatically generate reports, and update industry dynamics in real time.

• Exclusive customized modelsFor large enterprises, Cohere offers fully customized language model training services to ensure that the model meets the specific needs of the enterprise, such as accurate understanding and generation of industry terminology.

• core model::

  • Command-R: A new large-scale language model designed for enterprise-level retrieval-enhanced generation (RAG) and tool-use applications. With multi-language generation capabilities supporting 10 major business languages, output with explicit citations reduces the risk of hallucinations and presents more contextual information from the source material. Supports up to 128k of context with low latency, high throughput, and support for using tools (APIs) to automate complex tasks.
  • Embed and Rerank models: Support for more than 100 languages helps users derive answers from a large number of data sources, providing clear and accurate dialog in their native language, regardless of the language used.

Technical Advantages

• high performance: Cohere's model is deeply optimized based on the Transformer architecture, enabling efficient computation with fewer resources.
• high security: Offers private deployments and hybrid cloud solutions to ensure that organizations have complete control over data flows, meeting the needs of customers in industries that are highly sensitive to data privacy.
• Multi-Cloud Adaptability and Data Controllability: Cohere's products support both multi-cloud and local deployments, allowing organizations to choose the most appropriate deployment method based on their needs and security requirements.
• Strong customization capabilities: Focuses on providing customized solutions for enterprises, embedding powerful language models into their internal systems through APIs to meet their needs in a wide range of scenarios, including document search, text generation, and content summarization.
• interpretability: Improve AI interpretability and reduce the risk of AI hallucinations through techniques such as retrieval-enhanced generation.

Industry Collaboration and Client Practices

• Strategic Partners: Collaboration with Oracle (embedding its applications), Salesforce, Accenture, McKinsey, Fujitsu (Japanese LLM "Takane"), RBC (financial services "North" platform), LG (Korean language model), etc. North" platform for financial services), LG (Korean language model), and others..
• Customer Sample: Enterprise customers such as Oracle, Notion, Spotify, LivePerson, etc. use Cohere models for content generation, customer service conversations, document analysis, Q&A Agent construction, etc..
• Support for early-stage enterprises: Launches Startup Support Program, offering low-cost APIs, expert support and exposure to startups in Series B and below.